Skip to content

Latest commit

 

History

History
101 lines (58 loc) · 1.72 KB

2019-09-16.md

File metadata and controls

101 lines (58 loc) · 1.72 KB

nextstrain.org/flu and who.nextflu.org

Trevor, Barney, John, Jover, Thomas

The state of flu, prior to Barney's departure.

Fauna

vdb (virus database)

  • GISAID

e.g.

python vdb/flu_upload.py

Run weekly.

Manual click through searching, but can be an automated download soon.

tdb (titers database)

  • CDC emails seasonally-cumulative spreadsheets
  • Crick
  • VIDRL
  • NIID
  • (CCDC)

e.g.

python tdb/upload_all.py
python tdb/cdc_upload.py
python tdb/niid_upload.py

Data comes from

Data from all sources is normalized to CDC format.

Want an authoritative titers dataset.

No overlap between data sources.

Upload to tdb is an upsert.

nextstrain/seasonal-flu

nextstrain.org/flu

python batch.py --system batch [--lineages h3n2 h1n1pdm [...]]
nextstrain deploy auspice/*.json

Run weekly.

who.nextflu.org

python batch.py --version who
mv auspice-who/* ../nextflu/auspice/data/
cd ../nextflu/
bundle jekyll serve ...
aero deploy

Run weekly?

To become nextstrain.org/who-gisrs, we need to solve titers {visualization, surfacing} in Auspice.

Next steps

From Trevor:

  1. Improve deploy steps for who.nextflu.org by improving partial data updates
  2. Migrate to Nextstrain
  3. Migration away from Fauna

User goals for titers

"I have this serum. How good is it?"

Global view of antigenic diversity.

Maintenance goals

  • Automate more of the weekly updates

    • Updates for new locations?
    • Misspellings and other manual QC tasks

  • Build reliability

  • Make deploys and partial update deploys easier

  • Faster downloads from RethinkDB

    • Tuning? trs will try this.
    • Caching of static FASTAs on S3